Journal of Maps, 2009, 166-169
Integrated Magnetic and Global Positioning Satellite Mapping of the Firestone Reserve, Costa Rica
WARREN ROBERTS1 , DONALD A. MCFARLANE2 and KEITH CHRISTENSON2
1 Claremont
2 Keck
Colleges Libraries, Claremont CA 91711 USA; [email protected]
Science Center, The Claremont Colleges, 925 North Mills Avenue, Claremont CA 91711 USA;
Abstract
High-resolution topographic mapping of the Firestone Reserve, southwestern Costa Rica, has presented a number of challenges resulting from the poor-to-non-existent GPS satellite reception in steep
canyons and under tropical rainforest canopy. We have successfully employed magnetic survey techniques and data reduction software developed for cave mapping, integrated with sub-meter accuracy,
GPS-derived reference points in those areas where topography permits GPS reception. The resulting
topographic map is based on an x-y-z coordinate array averaging 284 data triplets per hectare and providing the level of resolution necessary to support microhabitat-level ecological studies.
(Received 22 nd January 2009; Revised 3 rd May 2009; Accepted 6th May 2009)
166
ISSN 1744-5647
doi:10.4113/jom.2009.1058
Journal of Maps, 2009, 166-169
1.
Roberts, W., McFarlane, D. A. & Christenson, K.
Introduction
The Firestone Center for Restoration Ecology (http://costarica.jsd.claremont.edu) operates a 60 ha ecological reserve in southwestern Costa Rica (latitude: 9.279 N; longitude
83.862 W) that is contiguous with the 330 ha Hacienda Barú National Wildlife Refuge
and adjacent units of the Path of the Tapir (Paso de la Danta) biological corridor. The
Firestone Reserve is characterized by high average relief (± 243 m/km), three precipitous ravines and at least 5 vegetation sub-habitats in what is nominally a Tropical Wet
Forest Holdridge Life Zone (Holdridge, 1967, Figure 1). At the time of acquisition in
March 2005, the best available map data was the 1:50,000 topographic national map series (Instituto Geográfico Nacional) with a 20 m contour interval. Proposed ecological
research on the reserve required reliable mapping of the topography at a resolution of 10
m or better, several orders of magnitude finer than could be extracted from the 1:50,000
maps. Since GPS reception was unavailable over approximately two thirds of the property due to tropical forest canopy cover and deep ravines the initial survey effort had
to use more traditional methods. We considered, and subsequently rejected, the use of
theodolites and laser total stations as being too expensive, too slow, and logistically impractical in very difficult terrain. Instead, we used a magnetic survey technique in wide
use for mapping cave systems throughout the World, which employs a high-precision
sighting compass and clinometer (Dasher, 1994). Reduction of the magnetic survey data
employed the cave survey program COMPASS (Fish, 2008). We then improved the accuracy by an order of magnitude by fixing 7 COMPASS survey stations using sub-meter
accuracy GPS and adjusting the remaining stations accordingly. The COMPASS data
was exported to ArcGIS (v9.2), integrated with a further series of GPS survey lines from
those areas with GPS reception, bfore a topographic surface was interpolated, and the
final map drawn.
2.
Methods and Results
Magnetic surveying employed a Suunto KB-14 sighting compass readable to ± 0.25
degrees, a Suunto PM5 sighting clinometer readable to ± 0.5 degrees, and a standard
surveyor’s tape read to ± 2.5 cm. The final data set comprised 1148 stations with 20.3
km of survey line. Mean survey shot length was 17.0 m. The data provided 46 loop closures, with loop lengths ranging from 94.3 to 2,521.5 m. Mean loop misclosure was 1.63
% (3 dimensional), and was distributed across the loop by the loop closure algorithm
in COMPASS (Fish, 2008). Finally, we used 7 COMPASS stations for which sub-meter
accuracy GPS data could be obtained (GPS real-time differential correction using a subscription service with OMNISTAR) as “fixed stations” in COMPASS, which corrects the
remaining dataset by treating segments between fixed stations as survey loops. Ac167
Journal of Maps, 2009, 166-169
Roberts, W., McFarlane, D. A. & Christenson, K.
Figure 1. Mapping a canyon on the Firestone Reserve.
curacy of the survey point of origin is ± 0.9 m horizontal and 2.1 m vertical, and the
nominal accuracy of any mapped point within the reserve is ± 1.09 m or better (all
errors quoted at 1 sigma).
The x-y-z COMPASS dataset was exported directly into ArcGIS (v9.2), and combined
with 15,978 sub-meter accuracy GPS survey points, features from paths and transects
in those areas of the reserve where good GPS reception was possible (landscape with
less obstructions overhead such as broadleaf foliage or outside canyon areas). The topographic surface was then modeled using the inverse distance weighted (IDW) interpolation tools in the ArcGIS Spatial Analyst module, applied to a 2 m resolution output
raster. Although available 1:50,000 topographic coverage uses the Ocotepeque 1935 datum, this projection is misaligned with commonly used GPS datums by as much as 260
m (Orvis, 2002) and precise corrections are not available due to some historical ambiguities in the datum origin. We therefore chose to project our data on the WGS-84 datum
making it directly compatible with consumer-grade GPS units and many current global
mapping datasets
168
Journal of Maps, 2009, 166-169
3.
Roberts, W., McFarlane, D. A. & Christenson, K.
Conclusion
The integration of magnetic cave surveying techniques with GPS-correction is a highly
successful and cost-effective technique for producing high resolution topographic GIS
datasets in a precipitous, tropical-forest canopy shielded environment.
Software
Magnetic survey data was processed using COMPASS (v5.08). GPS data was processed
with Pathfinder Office (v2.9). Subsequent data processing and cartography was performed with ESRIs ArcGIS (v9.2) and Spatial Analyst.
Acknowledgements
We thank our undergraduate survey teams Emily Haber, Elspeth Llewellyn, Christopher Wheeler, Jenny Aleman-Zometa, Kelly Janes, Callae Snively, Luanna Dobson, Sam
Scott, Emily Schultz, Keala Cummings, Alexandra Binder, and Alicia Hill for their efforts in the field. Real-time satellite correction services were provided by a subscription
through the Claremont Colleges Libraries (2007) and the Andrew Mellon Foundation
(2008). We are especially grateful to Carol Brandt, vice-president for International Programs, Pitzer College, for access to the reserve and for assistance in the field.
References
DASHER, G. R. (1994) On Station. A complete handbook for surveying and mapping caves, National
Speleological Society, Huntsville, 242 pp.
FISH, L. (2008) Compass [Online]. Available from: http://www.fountainware.com/compass/, [Last
accessed: 12 September, 2008].
HOLDRIDGE, L. R. (1967) Life zone ecology, Tropical Science Center, San Jose, Costa Rica.
ORVIS, K. H. (2002) GPS Locations and Costa Rican Topo Maps [Online]. Available from: http://web.
utk.edu/∼orvis, Unpublished manuscript, Department of Geography, University of Tennessee, [Last
accessed: 12 September, 2008].
169